At height 925051, AntPool mined two blocks. The stale block and header for 0000000000000000000199dc74f0c7d25adc618618ee00db80884b347253c264 can be found here.
The blocks are fairly similar in size and transaction count. Their header timestamps match.
Possibly, the templates for the two blocks were generated in different pool regions (e.g. US and EU) and two miners happened to find a valid nonce for them.
$ diff stale-coinbase-925051.json winning-coinbase-925051.json
2,4c2,4
< "in_active_chain": false,
< "txid": "d90685cda3a9b089b0a3d6767e97a2a3f0bff97bbcd4bfb3961f6f8f838024af",
< "hash": "ddc4800437f12adb71ba57564f9d087ae3c32e85452daa5b4fd55b116494b23c",
---
> "in_active_chain": true,
> "txid": "ee69df9fd46383c0270f10b1099fb2be7181d3c25598fba953d63c4ff7cae164",
> "hash": "cb5becf903406b5c6168daa0bfa7770853b057456784e2b917a62071d6cb9c0b",
6,8c6,8
< "size": 472,
< "vsize": 445,
< "weight": 1780,
---
> "size": 474,
> "vsize": 447,
> "weight": 1788,
12c12
< "coinbase": "037b1d0e1a4d696e656420627920416e74506f6f6c20ed006c02d60b80b001fabe6d6dcce60713ceeff291c999b9556fb739bf1ed548ef7e6600e6642b428d06a68d96100000000000000000009e9918d8000000000000",
---
> "coinbase": "037b1d0e1c4d696e656420627920416e74506f6f6c393730d2006302d60b312d02fabe6d6d5e131ec1ab2db5482bd3d4c83653a2e4b3a9f0223930e7f12982fc307be93c3610000000000000000000fa7b0f00000000000000",
32c32
< "value": 3.14547253,
---
> "value": 3.14549260,
46,48c46,48
< "asm": "OP_RETURN aa21a9ed700851f2aac740fc073b644b1af2084edf40d19e58a0cc4a66e621f82e28ea85",
< "desc": "raw(6a24aa21a9ed700851f2aac740fc073b644b1af2084edf40d19e58a0cc4a66e621f82e28ea85)#4a5r87p9",
< "hex": "6a24aa21a9ed700851f2aac740fc073b644b1af2084edf40d19e58a0cc4a66e621f82e28ea85",
---
> "asm": "OP_RETURN aa21a9ed915f1259d09a324ddc1552b8e1730279e9f0c8302bd2d85e487aceb52836cb2a",
> "desc": "raw(6a24aa21a9ed915f1259d09a324ddc1552b8e1730279e9f0c8302bd2d85e487aceb52836cb2a)#g9gn7u7y",
> "hex": "6a24aa21a9ed915f1259d09a324ddc1552b8e1730279e9f0c8302bd2d85e487aceb52836cb2a",
56,58c56,58
< "asm": "OP_RETURN 434f524501a37cf4faa0758b26dca666f3e36d42fa15cc01064e3ecda72cb7961caa4b541b1e322bcfe0b5a030",
< "desc": "raw(6a2d434f524501a37cf4faa0758b26dca666f3e36d42fa15cc01064e3ecda72cb7961caa4b541b1e322bcfe0b5a030)#djrqagdj",
< "hex": "6a2d434f524501a37cf4faa0758b26dca666f3e36d42fa15cc01064e3ecda72cb7961caa4b541b1e322bcfe0b5a030",
---
> "asm": "OP_RETURN 434f52450142fdeae88682a965939fee9b7b2bd5b99694ff644e3ecda72cb7961caa4b541b1e322bcfe0b5a030",
> "desc": "raw(6a2d434f52450142fdeae88682a965939fee9b7b2bd5b99694ff644e3ecda72cb7961caa4b541b1e322bcfe0b5a030)#genhdn6c",
> "hex": "6a2d434f52450142fdeae88682a965939fee9b7b2bd5b99694ff644e3ecda72cb7961caa4b541b1e322bcfe0b5a030",
93,95c93,97
< "hex": "010000000001010000000000000000000000000000000000000000000000000000000000000000ffffffff57037b1d0e1a4d696e656420627920416e74506f6f6c20ed006c02d60b80b001fabe6d6dcce60713ceeff291c999b9556fb739bf1ed548ef7e6600e6642b428d06a68d96100000000000000000009e9918d8000000000000ffffffff07220200000000000017a91442402a28dd61f2718a4b27ae72a4791d5bbdade787359cbf120000000017a9145249bdf2c131d43995cff42e8feee293f79297a8870000000000000000266a24aa21a9ed700851f2aac740fc073b644b1af2084edf40d19e58a0cc4a66e621f82e28ea8500000000000000002f6a2d434f524501a37cf4faa0758b26dca666f3e36d42fa15cc01064e3ecda72cb7961caa4b541b1e322bcfe0b5a0300000000000000000146a12455853415401000d130f0e0e0b041f1200130000000000000000296a27737973c358d62a6390f7cf4475c4c09496140a8e4c7dcfc3e558a300ef7bc72ce783d3aaad200000000000000000002b6a2952534b424c4f434b3a86499abc17e3b29a49658ba3335e54c353242303aa457af95a50b917007dc5950120000000000000000000000000000000000000000000000000000000000000000000000000",
< "blockhash": "0000000000000000000199dc74f0c7d25adc618618ee00db80884b347253c264",
< "confirmations": 0
---
> "hex": "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",
> "blockhash": "000000000000000000001a9d946d1c41cae21fb89a227b51371d32dcbad2f40d",
> "confirmations": 57,
> "time": 1764021928,
> "blocktime": 1764021928
Or they run different full nodes of different versions from which to serve block templates. This would also explain the “internal discrepancies” we’ve previously noticed: a pool sometimes mining ephemeral anchors and sometimes not, sometimes mining large opreturns and sometimes not.
ViaBTC, and now Antpool. Is this a new practice, or are we just observing this behaviour now?